colabfit/OMat24_train_aimd_from_PBE_3000_npt

OMat24 train aimd from PBE 3000 npt是一个大规模的开放材料数据集，基于密度泛函理论（DFT）计算。该数据集是OMat24的一部分，特别针对aimd-from-PBE-3000-npt训练划分。它包含了大量的分子结构和计算出的属性，分为训练集和验证集，每个集合都有基于不同结构生成策略的子数据集。数据集包含了6076290个唯一的分子配置和411540573个原子，涵盖了多种元素，并提供了能量、原子力和柯西应力等属性。

The aimd-from-PBE-3000-npt training split of OMat24 is a large-scale open dataset based on density functional theory (DFT) calculations. It is part of the OMat24 dataset, specifically tailored for the aimd-from-PBE-3000-npt training partition. It includes a vast number of molecular structures and calculated properties, divided into training and validation sets, each with sub-datasets based on different structure generation strategies. The dataset contains 6076290 unique molecular configurations and 411540573 atoms, covering a variety of elements, and provides properties such as energy, atomic forces, and Cauchy stress.